Implement StateAnalyzer

[robbisg]

This class must inherit from classical Analyzer function

Should it be used to analyze a sequence of ks or it is only a wrapper of sklearn clustering function?

• First option, well suited for analyzing metrics and so on, maybe it is too difficult to use it afterwards, difficult implementation
• Second option, smoother so that we use only sklearn + hmmlearn functions with fixed parameters, it can implement methods to get centroids and transform data in case we want to use a subsampled data for training and the full dataset for testing, maybe it is better also for RT, maybe. So that sequence of parameters can be used with AnalysisIterator.

I will implement second option.

• fit
• transform

[robbisg]

• Implement a StateAnalyzer, maybe it could be similar to Decoding
• When put metrics stuff
• Is there a way to optimize the stuff in AnalysisIterator
• When to use DownSamplers (Speed, Variance)